Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.kernel > #1710795 > unrolled thread

[PATCH 1/2] tracing: Fix kmemleak in tracing_map_array_free

Started byChunyu Hu <chuhu@redhat.com>
First post2017-08-14 12:20 +0200
Last post2017-08-14 14:10 +0200
Articles 2 — 2 participants

Back to article view | Back to linux.kernel


Contents

  [PATCH 1/2] tracing: Fix kmemleak in tracing_map_array_free Chunyu Hu <chuhu@redhat.com> - 2017-08-14 12:20 +0200
    Re: [PATCH 1/2] tracing: Fix kmemleak in tracing_map_array_free Chunyu Hu <chuhu.ncepu@gmail.com> - 2017-08-14 14:10 +0200

#1710795 — [PATCH 1/2] tracing: Fix kmemleak in tracing_map_array_free

FromChunyu Hu <chuhu@redhat.com>
Date2017-08-14 12:20 +0200
Subject[PATCH 1/2] tracing: Fix kmemleak in tracing_map_array_free
Message-ID<uekDw-2U0-7@gated-at.bofh.it>
kmemleak reported the below leak when I was doing clear of the hist
trigger. With this patch, the kmeamleak is gone.

unreferenced object 0xffff94322b63d760 (size 32):
  comm "bash", pid 1522, jiffies 4403687962 (age 2442.311s)
  hex dump (first 32 bytes):
    00 01 00 00 04 00 00 00 08 00 00 00 ff 00 00 00  ................
    10 00 00 00 00 00 00 00 80 a8 7a f2 31 94 ff ff  ..........z.1...
  backtrace:
    [<ffffffff9e96c27a>] kmemleak_alloc+0x4a/0xa0
    [<ffffffff9e424cba>] kmem_cache_alloc_trace+0xca/0x1d0
    [<ffffffff9e377736>] tracing_map_array_alloc+0x26/0x140
    [<ffffffff9e261be0>] kretprobe_trampoline+0x0/0x50
    [<ffffffff9e38b935>] create_hist_data+0x535/0x750
    [<ffffffff9e38bd47>] event_hist_trigger_func+0x1f7/0x420
    [<ffffffff9e38893d>] event_trigger_write+0xfd/0x1a0
    [<ffffffff9e44dfc7>] __vfs_write+0x37/0x170
    [<ffffffff9e44f552>] vfs_write+0xb2/0x1b0
    [<ffffffff9e450b85>] SyS_write+0x55/0xc0
    [<ffffffff9e203857>] do_syscall_64+0x67/0x150
    [<ffffffff9e977ce7>] return_from_SYSCALL_64+0x0/0x6a
    [<ffffffffffffffff>] 0xffffffffffffffff
unreferenced object 0xffff9431f27aa880 (size 128):
  comm "bash", pid 1522, jiffies 4403687962 (age 2442.311s)
  hex dump (first 32 bytes):
    00 00 8c 2a 32 94 ff ff 00 f0 8b 2a 32 94 ff ff  ...*2......*2...
    00 e0 8b 2a 32 94 ff ff 00 d0 8b 2a 32 94 ff ff  ...*2......*2...
  backtrace:
    [<ffffffff9e96c27a>] kmemleak_alloc+0x4a/0xa0
    [<ffffffff9e425348>] __kmalloc+0xe8/0x220
    [<ffffffff9e3777c1>] tracing_map_array_alloc+0xb1/0x140
    [<ffffffff9e261be0>] kretprobe_trampoline+0x0/0x50
    [<ffffffff9e38b935>] create_hist_data+0x535/0x750
    [<ffffffff9e38bd47>] event_hist_trigger_func+0x1f7/0x420
    [<ffffffff9e38893d>] event_trigger_write+0xfd/0x1a0
    [<ffffffff9e44dfc7>] __vfs_write+0x37/0x170
    [<ffffffff9e44f552>] vfs_write+0xb2/0x1b0
    [<ffffffff9e450b85>] SyS_write+0x55/0xc0
    [<ffffffff9e203857>] do_syscall_64+0x67/0x150
    [<ffffffff9e977ce7>] return_from_SYSCALL_64+0x0/0x6a
    [<ffffffffffffffff>] 0xffffffffffffffff

Signed-off-by: Chunyu Hu <chuhu@redhat.com>
---
 kernel/trace/tracing_map.c | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)

diff --git a/kernel/trace/tracing_map.c b/kernel/trace/tracing_map.c
index 0a689bb..305039b 100644
--- a/kernel/trace/tracing_map.c
+++ b/kernel/trace/tracing_map.c
@@ -221,16 +221,19 @@ void tracing_map_array_free(struct tracing_map_array *a)
 	if (!a)
 		return;
 
-	if (!a->pages) {
-		kfree(a);
-		return;
-	}
+	if (!a->pages)
+		goto free;
 
 	for (i = 0; i < a->n_pages; i++) {
 		if (!a->pages[i])
 			break;
 		free_page((unsigned long)a->pages[i]);
 	}
+
+	kfree(a->pages);
+
+ free:
+	kfree(a);
 }
 
 struct tracing_map_array *tracing_map_array_alloc(unsigned int n_elts,
-- 
1.8.3.1

[toc] | [next] | [standalone]


#1710884

FromChunyu Hu <chuhu.ncepu@gmail.com>
Date2017-08-14 14:10 +0200
Message-ID<uemlY-3Zn-21@gated-at.bofh.it>
In reply to#1710795
Resend to add linux-kernel cced. The steps I was using was:

echo hist:keys=irq > trigger
echo '!hist:keys=irq:vals=hitcount:sort=hitcount:size=2048' >  trigger

Then scan the kmem.
echo scan > /sys/kernel/debug/kmemleak

On 14 August 2017 at 18:18, Chunyu Hu <chuhu@redhat.com> wrote:
> kmemleak reported the below leak when I was doing clear of the hist
> trigger. With this patch, the kmeamleak is gone.
>
> unreferenced object 0xffff94322b63d760 (size 32):
>   comm "bash", pid 1522, jiffies 4403687962 (age 2442.311s)
>   hex dump (first 32 bytes):
>     00 01 00 00 04 00 00 00 08 00 00 00 ff 00 00 00  ................
>     10 00 00 00 00 00 00 00 80 a8 7a f2 31 94 ff ff  ..........z.1...
>   backtrace:
>     [<ffffffff9e96c27a>] kmemleak_alloc+0x4a/0xa0
>     [<ffffffff9e424cba>] kmem_cache_alloc_trace+0xca/0x1d0
>     [<ffffffff9e377736>] tracing_map_array_alloc+0x26/0x140
>     [<ffffffff9e261be0>] kretprobe_trampoline+0x0/0x50
>     [<ffffffff9e38b935>] create_hist_data+0x535/0x750
>     [<ffffffff9e38bd47>] event_hist_trigger_func+0x1f7/0x420
>     [<ffffffff9e38893d>] event_trigger_write+0xfd/0x1a0
>     [<ffffffff9e44dfc7>] __vfs_write+0x37/0x170
>     [<ffffffff9e44f552>] vfs_write+0xb2/0x1b0
>     [<ffffffff9e450b85>] SyS_write+0x55/0xc0
>     [<ffffffff9e203857>] do_syscall_64+0x67/0x150
>     [<ffffffff9e977ce7>] return_from_SYSCALL_64+0x0/0x6a
>     [<ffffffffffffffff>] 0xffffffffffffffff
> unreferenced object 0xffff9431f27aa880 (size 128):
>   comm "bash", pid 1522, jiffies 4403687962 (age 2442.311s)
>   hex dump (first 32 bytes):
>     00 00 8c 2a 32 94 ff ff 00 f0 8b 2a 32 94 ff ff  ...*2......*2...
>     00 e0 8b 2a 32 94 ff ff 00 d0 8b 2a 32 94 ff ff  ...*2......*2...
>   backtrace:
>     [<ffffffff9e96c27a>] kmemleak_alloc+0x4a/0xa0
>     [<ffffffff9e425348>] __kmalloc+0xe8/0x220
>     [<ffffffff9e3777c1>] tracing_map_array_alloc+0xb1/0x140
>     [<ffffffff9e261be0>] kretprobe_trampoline+0x0/0x50
>     [<ffffffff9e38b935>] create_hist_data+0x535/0x750
>     [<ffffffff9e38bd47>] event_hist_trigger_func+0x1f7/0x420
>     [<ffffffff9e38893d>] event_trigger_write+0xfd/0x1a0
>     [<ffffffff9e44dfc7>] __vfs_write+0x37/0x170
>     [<ffffffff9e44f552>] vfs_write+0xb2/0x1b0
>     [<ffffffff9e450b85>] SyS_write+0x55/0xc0
>     [<ffffffff9e203857>] do_syscall_64+0x67/0x150
>     [<ffffffff9e977ce7>] return_from_SYSCALL_64+0x0/0x6a
>     [<ffffffffffffffff>] 0xffffffffffffffff
>
> Signed-off-by: Chunyu Hu <chuhu@redhat.com>
> ---
>  kernel/trace/tracing_map.c | 11 +++++++----
>  1 file changed, 7 insertions(+), 4 deletions(-)
>
> diff --git a/kernel/trace/tracing_map.c b/kernel/trace/tracing_map.c
> index 0a689bb..305039b 100644
> --- a/kernel/trace/tracing_map.c
> +++ b/kernel/trace/tracing_map.c
> @@ -221,16 +221,19 @@ void tracing_map_array_free(struct tracing_map_array *a)
>         if (!a)
>                 return;
>
> -       if (!a->pages) {
> -               kfree(a);
> -               return;
> -       }
> +       if (!a->pages)
> +               goto free;
>
>         for (i = 0; i < a->n_pages; i++) {
>                 if (!a->pages[i])
>                         break;
>                 free_page((unsigned long)a->pages[i]);
>         }
> +
> +       kfree(a->pages);
> +
> + free:
> +       kfree(a);
>  }
>
>  struct tracing_map_array *tracing_map_array_alloc(unsigned int n_elts,
> --
> 1.8.3.1
>

[toc] | [prev] | [standalone]


Back to top | Article view | linux.kernel


csiph-web